sign in sign up

<<

8 the Orders and Rules of British Horseracing: Anticompetitive Agreements Or Good Governance of a Multi-Sided Sport?

8 the Orders and Rules of British Horseracing: Anticompetitive Agreements Or Good Governance of a Multi-Sided Sport?

C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 192 [192–216] 16.6.2009 11:52AM

8 The Orders and Rules of British Horseracing: anticompetitive agreements or good governance of a multi-sided sport?

Bruce Lyons

1. Introduction

In a sport that is linked with betting and its associated problems, British horseracing is internationally renowned for its quality, diversity and integrity. It is the second most popular sport in Britain, with nearly 6 million spectators watching it live and millions more watching at home, including a TV audience of 10 million for the Grand National each year. Racehorse owners spend £275 million for the excitement of watching their horses race.1 British horseracing also provides the punter’s favourite bet, leaving bookmakers with a gross win of over £1 billion, 10 per cent of which is put back into the sport. It is a key feature of the competitive analysis that three such diverse sets of consumers (spectators, owners and punters) buy into the same British horseracing product. This is the reason for calling it a multi-sided sport in the title to this chapter. Successful sports are built on strong governance, which is necessary to keep the competition exciting and free from corruption. Sports with weak, fragmented governance structures tend to lose public interest (e.g. boxing, wres- tling). The fact that British horseracing has had a unified governance structure for over 250 years is undoubtedly one of the contributing factors to its success. As one might expect over such a long period, a fairly lengthy set of regulations has been developed to govern the rules of individual races and the control of race, fixture and commercial rights. These rules and regulations are known as the ‘Orders and Rules of British Horseracing’, some of which were challenged by the UK Office of Fair Trading (OFT) as anticompetitive agreements under Chapter 1 of the Competition Act 1998 (equivalent to Article 81EC).

1 They recover only a third of this in prize money and sponsorship. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 193 [192–216] 16.6.2009 11:52AM

193 8. The Orders and Rules of British Horseracing

The competition issues under review in this chapter concern those Orders and Rules which had the effect of creating joint-selling rights in dealing with bookmakers, limiting the rights of individual racecourses to run fixtures whenever they want and restricting relative prize money across races.2 Such issues get to the tension between ensuring good governance and creating cartel-like restrictions. This is where economic analysis is necessary to determine which rules are necessary for good governance and which are not. A recurring theme in this chapter is that good governance in sport requires the appropriate treatment of externalities, which may be either positive (e.g. betting opportunities for bookmakers) or negative (e.g. inappropriate prize money in one race distorting incentives in another). A particularly important network externality is that the pleasure of horseracing for punters depends on the number and quality of horses in training (i.e. the number of owners and how much they spend).3 A famous idea in economics, known as the Coase Theorem, sets out the conditions under which bilateral bargaining can eliminate the inefficiencies associated with externalities. As we shall see, these conditions are unlikely to hold in the absence of certain of the Orders and Rules because property rights would be too fragmented and uncertain and transaction costs would be high. This case illustrates an important lesson for the practical application of competition policy. Restrictive agreements should not be considered exclu- sively from the perspective of their potential to create distortions, because many such agreements have a beneficial, efficiency-enhancing purpose. In such situations, it is necessary to consider the net benefit to consumers and act only against those agreements that are harmful or unnecessarily restrictive.4 More specifically, economists have become increasingly aware that there are important cases where more than one group of consumers gains benefit from the same product. While all such groups have similar interests in being able to buy into a high-quality product (or interface or platform or sport), the structure of prices (i.e. who contributes how much to funding the British horseracing product) can directly affect not only the distribution of benefits but also the design and quality of the product itself. In the case of horseracing,

2 Joint-selling of betting rights is similar to, but as we shall see importantly different from, the sale of media rights. For a short summary of competition issues in the sale of media rights in football, see Hatton et al. (2007). 3 A network externality arises when the value of the product to one consumer depends on the number of others. This is a cross-group externality when the beneﬁt to one group of consumers depends on the number of consumers in another distinct group. 4 A similar point is emphasised by Motta in Chapter 1 and Rey and Venit in Chapter 11. See also EAGCP (2005) for the economic approach to competition policy. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 194 [192–216] 16.6.2009 11:52AM

194 Cases in European Competition Policy: The Economic Analysis

a central issue is that a higher price charged to bookmakers for being able to take bets on British horseracing automatically feeds through to a lower price for owners and better quality racing for spectators. The reverse holds if bookmakers contribute less. Careful economic analysis is necessary to identify a benchmark optimum price structure and to understand the consequences of alternative structures for each group of consumers.5 With this last point in mind, section 2 discusses the preferences of the main groups of consumers. Section 3 introduces the roles of some key institutions of British horseracing. Section 4 summarises the statement of objections (known as a ‘Rule 14 Notice’) by the relevant UK competition authority – the Ofﬁce of Fair Trading. My own economic analysis is set out in section 5. Section 6 summarises the outcome of the case, both in terms of the modernisation of British horseracing and what has happened in various UK and European courts. It also provides a brief conclusion.

2. The consumers of British horseracing6

The aim of competition policy The principal objective of modern competition policy is to ensure that markets operate competitively in the interests of consumers. The OFT sums this up admirably in its mission as stated in successive annual reports and highlighted in its web home page headline: ‘Making markets work well for consumers.’7 The obvious first question to ask is, what are consumers buying? For most sports, it is rarely a single event such as an isolated race or self-standing football match that creates the thrill. The excitement is generated by a sporting competition which links results in different events. In the case of horseracing, there is no major league or knock-out cup, but horses develop their ratings to qualify for more highly rated events. Their form in one race also matters for handicapping in later races which may be run at any racecourse. Furthermore, the integrity of races run under a common governance structure underpins consumer confidence. This suggests a product definition of British horseracing, not an isolated race or day of racing, but an interlinked programme over the season. We return to this after considering consumers in more detail.

5 See Armstrong and Wright in Chapter 3 and Rochet in Chapter 7 for other examples of competition policy applied to two-sided markets. 6 Many of the ﬁgures used in this section can now be found in Deloitte (2006). 7 Emphasis in original. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 195 [192–216] 16.6.2009 11:52AM

195 8. The Orders and Rules of British Horseracing

What is the definition of a consumer? Generally speaking, we expect consumers to be the ones who spend money in order to enjoy the product, as distinct from producers who receive income in return for providing the product. This is a fairly straightforward definition for markets with a single category of consumer, but the concept is not so straightforward in a multi- sided market. When different groups of consumers are buying into the same product or ‘platform’ created with a large element of fixed costs, there are many possible ‘prices’ that could generate the same level of funding.8 However, this does not mean that all such sets of prices are equally good because they may affect the quality of the product itself. Put another way, the structure of prices matters. This theme is picked up in section 5. How should the interests of different groups of consumers be weighted? Given the difficulty of making interpersonal comparisons of utility, it seems reasonable to claim that any weights attached to groups of consumers should be non-negatively related to their total spending (i.e. financial contribution to the creation of the product) – those who spend more should not be considered less worthy than those who spend less on the product. For example, it would be inappropriate to consider the effect of a particular rule on bookmakers independently of its effect on owners and spectators.

Consumers and consumer preferences At the time of the case in 2003, there were around 9,000 racehorse owners with 13,000 horses in training. It cost around £17,000 p.a. to keep a horse in training and average prize money was only £6,000. The average price of a horse bought at two principal auction ﬁrms in Britain in 2005 was £28,000 (Deloitte, 2006). Although a very exceptional few horses go on to earn their owners a fortune at stud, most have a very much lower resale value.9 Racehorses are not attached to particular racecourses but kept and trained at trainers’ yards and raced at different courses across the country.

8 For example, newspapers typically sell advertising space to advertisers and content to readers. For some newspapers, the spend of each group is similar, but others vary enormously in the ﬁnancial contribution of advertisers and readers. At one extreme, ‘free newspapers’ get all their revenue from advertisers and at the other extreme readers pay for ‘free ads’ papers. Armstrong and Wright (see Chapter 3) refer to this as a ‘waterbed effect’ when a similar total revenue can be collected from different consumer groups in various alternative proportions. 9 The vast majority of owners have horses that fall into the latter category and can have no reasonable expectation of increasing their wealth through horseracing. Consequently, it is conservative to ignore the cost of horses and focus on training costs when considering the ﬁnancial contribution of owners. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 196 [192–216] 16.6.2009 11:53AM

196 Cases in European Competition Policy: The Economic Analysis

There were fifty-nine racecourses in Britain, including flat and national hunt (i.e. jumping) courses, at which spectators could watch horseracing. The highest quality of racing is on grass, but some courses have built all-weather tracks in recent years. The amount of racing on a grass course is strictly limited in any one period because it needs to recover from being cut up by galloping horses. All-weather tracks do not have this constraint. Most courses are individually owned, but there are three significant racecourse groups: Racecourse Holdings Trust (RHT) (thirteen courses), Northern Racing (nine courses) and Arena (seven courses including three with all-weather tracks). In 2003, there were 1,220 fixtures at which just over 8,000 races were run and a total of £94 million in prize money was on offer, 80 per cent of which went to winning owners. A very distinctive feature of horseracing is its link with the betting industry. Punters enjoy a bet on the races, either on-course, in off-course betting shops or increasingly on the internet. At the time of the case, in 2002/3, the gross win (i.e. revenue less payout) for bookmakers on British horseracing was £858 million, which was 42 per cent of their gross win from all betting.10 In recognition of this, the betting industry makes a substantial financial contribution to horseracing through the Levy (see below). This is used mainly for prize money to attract racehorse ownership, but also to fund integrity services and other support for horses and courses.11 Based on expenditure shares in 2000, the net financial contributions of the various consumer groups into British horseracing worked out at owners (via training fees, keep, vets’ fees, etc. net of prize money) 50 per cent, punters (via the Levy) 25 per cent, spectators (via racecourse attendance) 18 per cent, and sponsors andmedia(whoaremostlyinterestedinaccess to owners, punters and spectators) 7 per cent. In this context, although owners are less numerous than punters or spectators, they have a strong claim to being the most important consumer group.12

10 Football, ﬁxed-odds computer bets and other sources are now increasing the non-horseracing share, but British horseracing is still seen as a particularly attractive bet for punters. As recently as 1999, the British horseracing share of turnover (i.e. not quite equivalent to gross win) had been 70 per cent and latest ﬁgures for 2004/5 show it reduced to 38 per cent of gross win. However, this should be seen in the context of rapid growth in all betting so gross win on British horseracing had grown to £1.12 billion by 2004/5. 11 This is a very much smaller proportion of betting turnover than is contributed in other countries such as France, the US or Hong Kong (data from the International Federation of Horseracing Authorities). One consequence is that prize money is a lower proportion of training costs in Britain. Highest levels of prize money are available in countries with a state-sponsored tote or betting monopoly, which contrasts with the competitive British betting industry. 12 There are many other non-consumer interests, most of whom make their living out of horseracing, including trainers, jockeys, stable lads, transport drivers, vets, etc. These are sometimes referred to as ‘the industry’. Breeders merit a particular mention because horseraces are a crucial testing ground for their breeding skills, and true-run races provide important information in the continuing search for the best possible racehorse (known as ‘improvement of the breed’). C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 197 [192–216] 16.6.2009 11:53AM

197 8. The Orders and Rules of British Horseracing

While owners are very direct consumers, punters enjoy horseracing through the bookmaking intermediary. We adopt the standard convention in competition law and economics that direct customers (e.g. bookmakers) have an incentive to represent the interests of consumers who buy through them (e.g. punters), as long as they act competitively. A similar point may be made in relation to racecourses and spectators. This association between intermediate business customer and final consumer is not entirely unproble- matic but it is a useful starting point. Having identified the various consumers, it is useful to reflect on what they value from horseracing. Independent market research provides a guide to spectator utility functions. They value true-run races, competitive balance, seeing the fastest horses in the world, accurate information to help predict results, diversity of races and spectacle and glamour. Owners enjoy much the same attributes, but magnified by the vicarious pleasure of participating in a sport and added trips to the trainer’s yard to plan their horse’s campaign (i.e. season’s races). As an indication of punter preferences, there is unpublished econometric evidence (commissioned by bookmakers) that betting increases with TV cover- age, racing on a British racecourse, large fields (i.e. ten-plus runners), high prize money (i.e. higher quality of racing), turf (i.e. grass as distinct from all-weather tracks) and handicap races (i.e. more uniform spread of odds). Punters, spectators and owners all rely on racehorse form (results), built up across previous races at different racecourses, to predict results and plan campaigns. Thus, the same essential qualities are valued by different groups of consumers, though not necessarily with the same weightings. For example, owners generally prefer smaller race fields (and so a better chance of winning) than do punters (who dislike short-odds favourites). It is central to understanding this case that the core product each group of consumers is enjoying is the programme of races which costs half a billion pounds to put on each year. This is the ‘platform’ into which each of the consumer groups buys. The central competition issue is the extent to which an appropriately representative governing body is necessary to design horseracing for its various consumers and whether market mechanisms can be designed to deliver a more consumer-responsive product.

3. The institutions governing British horseracing

Different aspects of horseracing and its funding were coordinated by three key institutions: the Jockey Club, the British Horseracing Board (BHB) and the Horserace Betting Levy Board (HBLB). C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 198 [192–216] 16.6.2009 11:53AM

198 Cases in European Competition Policy: The Economic Analysis

The Jockey Club was formed in 1752 and governed most aspects of British horseracing until 1993. At the time of the case, it still governed sporting rules relating to the conduct of individual races (e.g. licensing of individuals, horses and courses; common standards across racecourses; anti-doping and main- taining integrity on- and off-course; veterinary care). Alongside its regulatory functions, it owned a portfolio of thirteen racecourses under the name of RHT.13 The Jockey Club was a not-for-profit organisation with all its net income going to support horseracing and the development of the thoroughbred as a breed. It had a charitable function in looking after retired and injured human and equine participants in horseracing.14 In 1993, BHB was formed to take over commercial activities and the sport’s overall coordination and organisation (e.g. funding, the fixture list, race planning, liaison with the betting industry and HBLB, marketing, strategic planning). BHB distributed no profits to its members, but invested all its income in the interests of the sport (racing, breeding and veterinary science). An important feature of BHB decision making was that it was representative of both regulatory and consumer interests, with the exception of the betting industry with which it negotiated at the HBLB.15 The OFT case mainly focused on those Orders and Rules of British Horseracing which enabled BHB to negotiate funding with bookmakers and determine fixtures and race planning. Owners, racecourses and others must accept these regulations if they want to take part in the races BHB puts on. The overall effect of these rules appears to give BHB considerable power over which courses can have fixtures at what times and the prize money for races run under different conditions, as well as in negotiating with bookmakers. In practice, the power to determine the fixture list is very severely constrained by implicit ‘grandfather rights’ of racecourses to keep their fixtures from one year to the next and by bookmakers through negotiations at the HBLB. Since 1961, the HBLB has been the statutory body (i.e. set up by government) which assesses and collects a monetary contribution (i.e. the Levy) from

13 The Racecourse Holdings Trust courses include four which host some of the most famous races in the world: Aintree (Grand National), Cheltenham (Gold Cup), Epsom (Derby and Oaks) and Newmarket (2000 and 1000 Guineas). 14 This was the governance structure at the time of the case but it has since changed. All remaining regulatory functions of the Jockey Club have been brought together with BHB to create the British Horseracing Authority (BHA). For the economic role set out in this chapter, the BHA has effectively the same role as BHB. 15 BHB policy was determined by a thirteen-person board of directors, appointed by the following: Jockey Club (three), Industry Committee (two), Racecourse Association (RCA) (two), Racehorse Owners Association (two), Thoroughbred Breeders Association (one) and the board itself appointed a chairman, another independent director and the chief executive. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 199 [192–216] 16.6.2009 11:53AM

199 8. The Orders and Rules of British Horseracing

bookmakers in support of the attractive betting opportunity that British horseracing provides.16 It determines the size of the Levy in conjunction with incentives to provide a fixture list that suits the bookmakers. Under the mediation of an independent HBLB chair, representatives of BHB negotiate with representatives of the Bookmakers’ Committee (a committee of leading bookmakers and bookmakers’ associations). Formally, the Bookmakers’ Committee makes an annual proposal to the HBLB with a view to BHB agreeing to provide a fixture list that suits bookmaker interests (i.e. regular races of sufficient quality spread out across the week).17 The UK government’s role in horseracing and the betting industry appears anomalous, so it was receptive to proposals that would allow disengagement without harming the sport.18 What was needed was a mechanism for BHB to prevent bookmakers from freeriding on British horseracing and they came up with a creative proposal. Since bookmakers need precise information on runners, riders and associated pre-race data in order to take bets on races, BHB proposed to sell this data to bookmakers as a replacement for the Levy.19 This new ‘commercial mechanism’ was phased in over a period of five years from 2001, with the amount charged for supplying pre-race data being set against the Levy. The intention was that this would allow the HBLB to be abolished in 2005 (i.e. once the robustness of this commercial mechanism had been tested). The price for the data was on exactly the same basis as the Levy (i.e. a proportion of each bookmaker’s gross win on British horseracing) but this proportion rose under the commercial mechanism and stabilised at around 10 per cent of the gross win on British horseracing. The increasing importance of the Levy at the time of the case (2000–4) is shown in Figure 8.1.20

16 In 1928, the Tote was established by Act of Parliament with a statutory monopoly in non-fixed odds pool betting and the requirement to distribute its profits for ‘purposes conducive to the improvement of breeds of horses or the sport of horseracing’. HBLB was established in anticipation of the legalisation of high-street betting shops in 1962. It brought together contributions from the fixed-odds high-street bookmakers and Tote profits. 17 If there is a failure to agree, the Secretary of State (i.e. a government minister) is required to decide and this is usually a sufficient threat for agreement to be reached. 18 Having said this, there is far less government intervention in British horseracing than in almost any other jurisdiction. The UK government was also looking for ways to privatise the Tote without compromising the funding of British horseracing. A private sale at a discount to the Racing Trust (a consortium including BHB, the Jockey Club and RCA) was blocked by the European Commission on state-aid grounds (competition policy has a pervasive effect!) and the proposal at the time of writing is to sell the Tote on the open market. 19 At the time of the case, BHB’s property (database) rights on pre-race data, which are essential if it is to be sold at an appropriate price, had been confirmed in the High Court (2001) when challenged by William Hill (a leading bookmaker). This decision was overturned by the Court of Appeal in 2005. 20 The increase is substantially attributable to a change in betting tax in the 2001 budget away from a 9 per cent tax on bets taken (which punters saw as directly reducing their odds, i.e. raising the price of a bet) in C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 200 [192–216] 16.6.2009 11:53AM

200 Cases in European Competition Policy: The Economic Analysis

Horserace betting Levy yield 120

100

80

£m 60

40

20

0

1995/6 1996/7 1997/8 1998/9 2000/1 2001/2 2002/3 2003/4 2004/5 1999/2000 Figure 8.1 Trend in betting’s contribution to British horseracing Source: HBLB website.

Part of the Levy goes to fund integrity services and other support for horseracing and race courses, and around 60 per cent goes into the prize fund. £100 million p.a. provides ample incentive for a wide range of interests to use whatever legal means might be available (including competition law and IPR law) to increase their slice of the cake even if, as we shall see, their actions if successful would lead to a product less attractive to consumers (i.e. a smaller cake may be baked). Bookmakers lobbied hard with the OFT and divisions arose between racecourses.

4. The OFT Statement of Objections

In summer 2000, BHB and the Jockey Club notified to the OFT the Orders and Rules and other agreements which govern British horseracing in order to gain clearance that they did not infringe competition law. Eventually, in April 2003, the OFT issued what is known as a Rule 14 Notice, setting out a Statement of Objections under Chapter 1 of the Competition Act (i.e. equivalent to EC Article 81). The full Rule 14 Notice contains confidential information, but a non-confidential informal summary was issued at the same time:

favour of a less distortionary 15 per cent tax on a bookmaker’s gross win (i.e. the difference between bets taken and winnings paid out). In addition to this tax-related increase in betting, the proposed price of pre-race data now came from BHB, not the Bookmakers’ Committee, which improved BHB’s relative bargaining power. The formal Levy adjusted to the price of pre-race data. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 201 [192–216] 16.6.2009 11:53AM

201 8. The Orders and Rules of British Horseracing

‘In essence, the OFT has taken the preliminary view that certain Orders and Rules infringe the [Competition] Act on the grounds that they have the combined effect of: limiting the freedom of racecourses to organise their racing, in particular by fixing how often and at what times they stage races and the type of race they stage; fixing the amounts that racecourses must offer owners to enter their horses in a race; and monopolising the supply of race and runners data to bookmakers by foreclosing competition from alternative suppliers.’21 The influence of the Orders and Rules in regulating race fixtures and types, prize money and the sale of pre-race data was not in dispute. The dispute was over who were the relevant consumers and the impact on consumer welfare. The OFT accepted that the application of competition rules to a sport must be sensitive to the distinctive characteristics of that sport so that ‘certain sporting rules will not infringe the Chapter 1 prohibition if they are essential to enable the sport to operate’ (OFT654, 2003, para. 2.1). However, it considered that some consequences were hard-core market sharing and price fixing and were not essential to achieving the sporting objectives. It made no serious attempt to consider the implications of striking out the apparently offending Orders and Rules (i.e. the counterfactual). The OFT did attempt to define a relevant market for each of the three bullet points above. It then proceeded by identifying the producers and customers in each of its three defined separate markets and drew the implication that there should be no rules to interfere with the freedom of action of individual producers in each market. In section 5, we argue that a fundamental problem with the OFT’s original analysis is that it is inappropriate to consider horseracing as the sum of three independent markets.22 This commits the serious mistake of applying one-sided logic to the analysis of a multi-sided product. Before developing this point, however, we set out the OFT position on the three ‘markets’ in a little more detail.

The ‘market for fixtures and programmes’ A fixture is an event at a racecourse that contains a programme of five or six races on a particular date. The programme refers to the number of races at a

21 OFT654, 2003, para. 1.5. 22 More precisely, the OFT position was that there were many other component markets (e.g. the market for racehorse training), but these three were the ones for which it thought the Orders and Rules were anticompetitive. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 202 [192–216] 16.6.2009 11:53AM

202 Cases in European Competition Policy: The Economic Analysis

fixture, the conditions for each race (e.g. handicap versus equal weight carried by horses; flat racing versus national hunt over fences) and start times. The OFT view was that racecourses are the natural producers of fixtures and so should be allowed to organise them whenever they individually perceive there to be the demand. New racecourses should also be free to enter, and a healthy market should see weaker racecourses exit. This does not happen. The number of racecourses had been stable at fifty-nine with neither entry nor exit for over half a century, which suggested that the licensing system was over-restrictive. Basically, there was no mechanism to reward efficient racecourses and penalise those that were unattractive to consumers. The rules that were deemed anticompetitive included those preventing racecourses, without BHB approval, from introducing new fixtures even when the course could physically take more, or holding a fixture under the same code (i.e. flat versus jumps) at the same time as another racecourse if located within fifty miles. The OFT also argued that each racecourse should be allowed to determine the content of a day’s racing, including running however many races its track could safely run, at whatever time and with whatever conditions on the race that it wanted.

The ‘market for British racing opportunities’ The OFT saw this as a separate market in which the racecourses supply opportunities for owners (i.e. the customers) to race their horses. The price of such opportunities has two main components: the stakes owners must pay in order to enter a race and the prize money they might receive if their horse wins or is highly placed. Various Orders and Rules restrict the range of stake money and, quantitatively much more important, regulate prize money. Prize money regulation has several dimensions, including minimum prize money for a race, relative prize money across different classes of race (e.g. according to the quality rating of horses) and amount of place money relative to the winner’s prize. These were seen as classic price ﬁxing.

The ‘market for pre-race data’ Before bookmakers can take money on a race, they and their punters need to know which horses are racing, the handicap weight being carried and pre- ferably also who is riding each horse. Further important information includes C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 203 [192–216] 16.6.2009 11:53AM

203 8. The Orders and Rules of British Horseracing

each horse’s ‘form’ (i.e. recent results) and other details that may affect judgements as to each horse’s chances of winning. Such information is meticulously collated for BHB by a private company called Weatherbys.23 This information is provided free to newspapers to publicise race fixtures. However, from 2001 it was sold to bookmakers by BHB as part of the new ‘commercial mechanism’ to replace the Levy. The OFT considered that BHB had created an effective monopoly in the supply of this data because the notification requirements in the Orders and Rules mean that owners and trainers are obliged to notify Weatherbys of which horses they propose to run in which races. If a race is oversubscribed, Weatherbys follows fixed priority (based mainly on a horse’s rating) and balloting rules and notifies owners whose horses are excluded from being able to run in that race. The OFT claimed that these rules effectively excluded racecourses or third parties from being able to sell their own pre-race data to bookmakers and so prevented competition in the supply of data.

Some problems with the OFT’s Statement of Objections The OFT analysis identified three apparently independent ‘markets’. From this perspective, it could see no connection, for example, between owners as suppliers of racehorses for the ‘market for fixtures and programmes’ and owners as consumers of the ‘market for British racing opportunities’!It provided no view of how the organisation and quality of British horseracing would evolve in the absence of the offending Orders and Rules (i.e. the counterfactual). There was no recognition of the pre-race data as a legitimate property right over the overall British horseracing product, its role in the commercial mechanism, or the consequences if several firms competed in the supply of pre-race data. Finally, there was no appreciation of the beneficial effects of some of the key rules under challenge. Of course, the adoption of a purely prosecutorial position might be seen as a legitimate strategy for a competition authority operating in a court system, in order to encourage the defendants to articulate and evaluate what the benefits are. However, this is not appropriate in a UK or EU-type administrative system where the authority acts as judge and jury.

23 This is a family company that has a centuries-long relationship with horseracing. Since 1770 it has traced the lineage of every racehorse and so has a crucial role in horse breeding. It also performs banking functions for horseracing (e.g. related to stake and prize money) and prints racecards for the courses. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 204 [192–216] 16.6.2009 11:53AM

204 Cases in European Competition Policy: The Economic Analysis

5. Economic analysis of the offending Orders and Rules

Several of the Orders and Rules challenged under the ‘market for ﬁxtures and programmes’ were indeed overly restrictive, but others have essential incentive and efﬁciency properties. Their prohibition would have fundamentally devalued British horseracing. Economically the most important relate to the sale of pre-race data, so this is the main emphasis of this chapter, though we return to the role of other Orders challenged by the OFT at the end of this section.

The betting externality and optimal bookmaker contribution to British horseracing24 In this section, we provide a simple model designed to capture the essential features of the relationship between horseracing and bookmakers, in particular, the role of prize money in raising the quality of horseracing enjoyed by its various consumers. The model allows us to conceptualise the contribution from bookmakers that would maximise the beneﬁt of all consumers. It also provides a benchmark from which to conduct a counterfactual analysis of what would happen if certain Orders and Rules were prohibited. For simpli- city, we focus on racehorse owners and bookmakers, leaving racecourses/ spectators and others in the background. However, a substantial proportion of the Levy goes towards providing integrity services, vet science and racecourse investment, for which very similar arguments apply.

A simple model of the betting externality25 Bookmakers make profits on bets taken on races and these bets increase with the number (and quality) of racehorses. Apart from providing more races on which to bet, punters bet more on higher-profile races, well-known horses and races with larger ‘fields’ (i.e. more horses per race).26 To capture this, we write

24 This section is more technical than others in this chapter, but the text has been written with a view to it being self-explanatory for those who prefer to avoid equations. 25 The formal approach taken here is similar to that in the recent literature on two-sided markets. See the references given in Chapters 3 and 7. It differs in that the two-sided market literature has one or more platform owners who set out to maximise proﬁts. The literature then examines the implications for pricing structures. In the current context, it is more appropriate to model negotiations between different groups with similar interests in developing a common product or ‘platform’. Nevertheless, the key ideas of multiple groups of consumers and appropriate pricing structures in the presence of positive externalities are essentially the same, and so too are the essential insights. 26 These assumptions reﬂect research by leading bookmakers. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 205 [192–216] 16.6.2009 11:53AM

205 8. The Orders and Rules of British Horseracing

the ‘gross win’ for bookmakers as B(n) where n is the number of racehorses in training and available to race. The more horses in racing, the greater the benefit to bookmakers: B0(n)=β(n)>0.27 If bookmakers could not, for any reason, take bets on British horseracing, they could offer punters alternative betting opportunities (e.g. Italian horseracing, computer-generated horseracing, dog racing) on which they could expect a gross win of p. The ‘betting externality’ is the positive contribution that is specific to British horseracing: BðnÞp. If bookmakers also contribute P ≥ 0 to the prize fund for races, the net benefit to bookmakers is:28 p ¼ BðnÞp P (1)

Racehorse owners gain enjoyment out of seeing their horses race, and particularly seeing them win. We write the (inverse) demand for racehorses in training as v(n) with v0(n) < 0. Racehorse ownership is expensive, both the original purchase of a horse and the costs of training and keep. We represent the annualised cost as c per horse. Part of the cost is mitigated, at least in expectation terms, by the opportunity to win prize money. On average, each racehorse can expect to win P/n out of the bookmakers’ contribution to the prize fund, so the net beneﬁt of owning the nth horse is u(n)=v(n) − c + P/n.29 The equilibrium number of horses in training is determined by where ownership of the marginal horse brings pleasure and expected prize money just equal to the costs of ownership: vðnÞþP=n ¼ c (2)

Lower costs or higher prize money encourage new owners into racing and existing owners to keep more and better quality horses in training. In particular:30 dn=dP ¼ 1=½ðP=nÞnv0ðnÞ 40 (3)

To give an idea of the size of this prize-fund effect, the increased Levy shown in Figure 8.1 was associated with a growth (2002–5) of: total prize

27 A prime denotes a partial derivative. There is diminishing marginal benefitsoβ0(n)<0. 28 In practice, the prize fund depends on the gross win, so we could write P = tβ(n) where t = 10 per cent is the size of the Levy. It turns out that, because a lump-sum Levy has the same economic effect as a ‘profits tax’, we do not need to be concerned about the precise specification of the prize fund. 29 Owners also contribute stake money for races, but this goes to the winner, so it does not affect the average net benefit of ownership. Average prize money per horse is over £6,000 p.a., though the distribution is highly skewed, with the top ten horses winning £300,000–£800,000 each and many winning nothing. There is no doubt, however, that most owners of the latter still have a positive expectation of winning! 30 This is derived by totally differentiating (2) and rearranging. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 206 [192–216] 16.6.2009 11:53AM

206 Cases in European Competition Policy: The Economic Analysis

money (net of owner contributions) of 20 per cent; and horses in training, runners, races and ﬁxtures, each of between 11 per cent and 13 per cent. The long-term effects are likely to be larger. There is also some evidence of an increase in the quality of British-trained horses.31 Since there are n horses, the beneﬁt to owners as a group is: UðnÞ¼VðnÞnc þ P (4)

where V(n) is the gross utility of racehorse ownership.32 Even without a prize fund from the bookmakers, a lower level of horseracing would exist. We define n as the number (and quality) of horses that would be in training if P = 0 and the consequent benefits to owners as U ¼ VðnÞnc .33 Before developing implications for negotiations between bookmakers and owners, we establish the socially optimal outcome for this two-sided market with a one-sided externality (i.e. bookmakers benefit from racehorse ownership but not directly vice versa). Consider what the optimal size (and quality) of British horseracing, and so also the optimal prize fund, would be if we were interested in maximising the joint benefit of owners and bookmakers. Adding (1) and (4): ωðnÞ¼VðnÞnc þ BðnÞ (5)

This implies that the socially optimal number (and quality) of racehorses is n*, which is found by solving: vðnÞþβðnÞ¼c (6)

In words, the combined marginal beneﬁts to owners and bookmakers of an extra horse should equal the marginal cost.34 Of course, the optimal design of horseracing is a lot more complex than a single number, n*, and this is discussed brieﬂy below, but for now it will serve as a simple measure of

31 It is very difficult to measure quality but one international comparison casts light at least on the best racehorses. If we calculate the difference between prize money won abroad by UK-trained horses and prize money won in the UK by foreign-trained horses, and compare the 2004/6 average with the 2002/3 average, this also grew by 13 per cent. All figures in this para. are taken from the BHB annual report (2005) supplemented by the BHB (2006) report on ‘Fixtures, Races and Prize Money’ available on its website. As the Levy and prize money have reduced more recently, there is evidence of these effects reversing (BHA, 2007). 32 Formally, V(n) is the integral of v(n) over the n horses. Non-marginal owners obtain positive consumer surplus which is measured by U(n). 33 Formally, from (2), n solves vðnÞ¼c. 34 Note that v(n*) − c < 0 for (6) to hold which explains why extra prize money is necessary to bridge the deficit for owners. Prize money has no direct effect on ω because for every £1 that owners gain, bookmakers lose the same. However, there is a crucial indirect effect because a higher prize fund means more (and better quality) racehorses, which benefits both owners and bookmakers. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 207 [192–216] 16.6.2009 11:53AM

207 8. The Orders and Rules of British Horseracing

the overall quality of racing and facilitate discussion of the fundamental economic issues.

The Coase Theorem does not hold for horseracing The Coase Theorem states that if there are clearly defined property rights and in the absence of transaction costs, free negotiations should eliminate any distortions due to an externality and so result in the optimal outcome (in this case the optimal number and quality of racehorses).35 The intuition is that it is mutually beneficial for bookmakers and owners to get round a table to agree on n*, which maximises joint utilities, and make a payment to compensate owners for providing more races than they would otherwise do. This transfer payment would only affect the distribution of utility between the parties (and ensure that each gets at least their reservation utilities U and p). The first problem with this is that it is not practical to negotiate directly over n, which is determined by the decisions of thousands of owners. Even if it was possible to negotiate with so many dispersed individuals, it is simply not feasible to negotiate in a way that does not compromise the integrity of racing – it is not credible to expect true-run races when bookmakers are seen to be ‘bribing’ owners (even if for the best of motives). Owners have to negotiate through an intermediary with the power to act on their behalf (e.g. BHB). However, even then, the intermediary cannot force owners to keep n* horses in training. This can only work indirectly through a prize fund attracting ownership. The consequences are developed below, particularly focusing on the fact that the quality of racing is not independent of the distribution of bargaining power between owners and bookmakers. An additional complication is that negotiations simultaneously take place over the fixture list and timing of races with a view to maximising the positive betting externality (e.g. by having fixtures throughout the week, every week and spread out during the day, so that betting opportunities are enhanced). This means that the negotiating intermediary must be able to deliver the fixture list as well as ownership incentives. This requires an intermediary who can also represent racecourses. In order for the intermediary to be able to represent these interests it must have a secure ‘property right’ to be able to exclude bookmakers from taking off-course bets on British horseraces; otherwise, it would have no bargaining power. The Levy system acted equivalently and the ownership of pre-race

35 The classic references are Coase (1960) and Williamson (1985), but the basic principles are set out in any good intermediate microeconomics textbook. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 208 [192–216] 16.6.2009 11:53AM

208 Cases in European Competition Policy: The Economic Analysis

database rights was intended to enable commercial rights without the neces- sity of government intervention through the HBLB. At the time of the case, this key property right appeared to be in place and BHB had the ultimate right to exclude bookmakers from taking bets on their races.36

Collective negotiations between BHB and the bookmakers In order to understand the problems raised by the OFT’s ‘preliminary view’ that multiple suppliers should be free to provide competing databases, I set the scene by considering the status-quo negotiations (i.e. as of 2003), with BHB selling the right to take bets and the Bookmakers’ Committee buying this right. The outcome of such bargaining is hard to predict precisely, but it is instructive to consider the extremes, such as when one party can make a take- it-or-leave-it offer to the other. If bookmakers had all the bargaining power, maximising (1) gives the implicit number of racehorses, nB, that they would want to support:37 ½vðnBÞþnBv0ðnBÞ þ βðnBÞ¼c (7)

Equation (7) differs from (6) in that marginal revenue (i.e. the term in square brackets) replaces the benefit of the marginal owner. The second element of the term in square brackets is negative which means that nB < n*. In other words, from the perspective of all consumers, there would be too little horseracing (of too low a quality) taking place. Thus, even with all the bargaining power, bookmakers want to contribute to a limited prize fund in order to bring more horses into racing and so generate a higher gross win on the associated bets.38 However, there would be too little horseracing (and of a quality that is too low). At the other extreme, if BHB (acting on behalf of owners and spectators) had all the bargaining power, the bookmakers would still have the option to stop taking bets on British horseracing, and they would exercise this option if BHB demanded too much. This ‘quit option’ is the key to considering what would happen if owners could dictate terms to bookmakers. BHB could force net benefit down to zero, so from (1) it could press for Pmax ¼ BðnOÞp.Wecanfind the number of racehorses this would bring into training by substituting into (2): BðnOÞp vðnOÞþ ¼ c (8) nO

36 See footnote 19. 37 This expression follows after the substitution of (2) and (3). 38 The prize fund would be positive as long as nB4n, which is certainly the case for British horseracing. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 209 [192–216] 16.6.2009 11:53AM

209 8. The Orders and Rules of British Horseracing

The first term in square brackets is the average benefit to bookmakers, which exceeds the marginal benefit so, by comparison with (6), this tends to result in too many racehorses. However, the p term works in the opposite direction, so the overall effect depends on the outside opportunities of bookmakers and may result in too little prize money.39 Combined with our result when bookmakers can dictate terms, we find: n5nB5n nO (9)

Thus, whatever the allocation of bargaining power, bookmakers will want to invest a positive amount in racing, but if their bargaining power is strong, they will invest too little. The efficiency of owners having maximum bargaining power is less clear. As long as there is collective negotiation on both sides, the bookmakers’ contribution is likely to lead to between nB and nO. The actual outcome is then determined by relative negotiating skills and the institutional setting of the bargaining process. We cannot say that this outcome will be precisely the optimal n*. However, this analysis does give reason to believe that the outcome is at least in the right ball park when there is collective negotiation on both sides; BHB is able to exclude bookmakers from taking bets on British horseracing and bookmakers are free to walk away. It is helpful to illustrate issues in the following subsection with some rough orders of magnitude to set the bounds of negotiation. The cost of filling daily fixture gaps in the racing calendar (i.e. fixtures that would not happen in the absence of bookmaker funding) might be achieved for around £20 million, though this would not bring about high-quality racing. At the other extreme, while bookmakers take a gross win of over £1 billion p.a. on British horseracing, there would undoubtedly be considerable substitution to other bets if punters were prevented from betting on this. The extent of possible substitution is possibly known by bookmakers, but if it is, it is a closely guarded commercial secret. An educated guess might put the maximum bookmakers would con- ceivably pay at £200 million, a figure which will serve for illustrative purposes. The central conclusion of this section is that due to the positive externality created for the betting industry by British horseracing, and the dispersed decisions by owners, an efficient market requires bookmakers to return a substantial share of their profits to invest in prize money and so enhance the size and quality of racing: n5n.40 This contribution should be higher than

39 With a sufficiently strong outside option nO < n* and nB ≤ nO. Inequality (9) assumes the bookmakers’ outside options are more limited. 40 This efficiency argument is quite separate from any moral case that beneficiaries should pay for positive externalities. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 210 [192–216] 16.6.2009 11:53AM

210 Cases in European Competition Policy: The Economic Analysis

what bookmakers would provide on their own initiative: nB < n*. It is against this broad background that we can evaluate the counterfactual, i.e. the consequences of the OFT proposal to require disaggregated bargaining on the part of horseracing interests, with each racecourse negotiating separately.

Implications of alternative institutional arrangements for negotiating over the betting externality Competition policy practitioners should carefully consider the economic consequences of their actions. They need to be particularly careful when there are beneﬁcial effects of apparently anticompetitive restrictions. Consider the following alternative negotiating structures which represent the range of possibilities that might have arisen depending on the precise outcome of the OFT’s investigation.41 The ﬁrst four assume the continued existence of clear property rights that permit BHB (or, for cases 3 and 4, individual racecourses) to exclude bookmakers from taking bets unless prior agreement has been reached: 1. BHB negotiates with collective bookmakers. 2. BHB negotiates with individual bookmakers. 3. Individual courses negotiate with collective bookmakers. 4. Individual courses negotiate with individual bookmakers. 5. No property rights for BHB (or racecourses) to exclude bets on British horseracing.

1. BHB negotiates with collective bookmakers This represents the status quo as already discussed. Both the racing side and the betting side act collectively. If negotiations between two parties are balanced, we expect that the beneﬁts will be shared roughly equally. Equal monetary shares between the maximum and minimum contributions by bookmakers would suggest a Levy (or sale of pre-race data) of around £110 million (= ½[£20 m + £200 m]).

2. BHB negotiates with individual bookmakers Suppose the OFT had focused only on the collective bargaining of bookmakers, instead of the collective bargaining only of the racing side in its

41 The OFT was silent on whether bookmakers would continue to be able to negotiate collectively. We separate out the consequences of ‘competitive provision of pre-race data’ because it would have distinctive economic effects. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 211 [192–216] 16.6.2009 11:53AM

211 8. The Orders and Rules of British Horseracing

Rule 14 Notice, and prohibited the Bookmakers’ Committee. This would have allowed BHB to play off one bookmaker against another in order to extract the full surplus of £200 million. It might even go further and offer exclusive betting rights to one bookmaker, who would then be able to act as a monopo- list (e.g. offering worse odds to punters). This might bring in even more money for the racing side, but it would do so by exploiting punters.42

3. Individual courses negotiate with collective bookmakers This would have been the most likely outcome if the OFT had pursued its ‘preliminary finding’ that racecourses should independently negotiate their fixtures and funding from bookmakers. If no single racecourse group was essential to providing a full set of fixtures required by the bookmakers, the latter could play one off against the other to drive payment down to the minimum £20 million. In effect, the bookmakers would design British horseracing with no account taken of other consumer interests.43

4. Individual courses negotiate with individual bookmakers Large bookmakers would continue to fund certain ﬁxtures but only as long as they could prevent rivals from freeriding on them. Bookmaker-funded ﬁx- tures would be subject to exclusivity clauses, thus foreclosing smaller bookmakers. The effects would be to increase horizontal concentration amongst bookmakers and vertical integration with racecourses. As with case 2, punters would be faced with more expensive bets. With or without exclusivity, this set of negotiations would create enormous coordination problems in the design of the racing calendar because horses are not tied to courses in the UK. The problems created by the absence of a single governing body coordinating and creating sporting competition are vividly illustrated by the loss of consumer interest in boxing when sanctioning bodies proliferated.

5. No property rights for BHB (or racecourses) to exclude bets on British horseracing The OFT’s preliminary view was that the supply of pre-race data was being monopolised by BHB and others should be given the right to sell it. Suppose that competitive supply could be established. This would cost around £5 million p.a. for each provider. If a genuinely contestable market could be

42 At present, British horseraces provide a ‘low price’ bet compared with other betting opportunities, i.e. an odds-weighted bet on each horse in a race would on average return a punter a higher percentage of his outlay than on alternative bets such as racing from another country, dog racing, etc. Of course, this percentage is less than 100 per cent, so the bookmakers earn a proﬁt. 43 In terms of our formal analysis, this would result in nB. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 212 [192–216] 16.6.2009 11:53AM

212 Cases in European Competition Policy: The Economic Analysis

created, bookmakers could purchase at cost price. If acting collectively, they would buy the data for £5 million, but none of this would contribute to the fixture programme for which they could then negotiate and fund as in case 3. If acting individually, they would pay for the data, but no doubt freeride on fixtures with an outcome no better than in case 4. Of course, the idea of a competitive market in pre-race data provision completely misses the point of a commercial mechanism designed to replace the Levy. At best (i.e. if bookmakers act as a cartel) it would arrive at the unbalanced outcome of case 3, and at worst (i.e. if bookmakers act individually) it would make everybody worse off. In practice, competitive data supply would be even worse than in cases 3 or 4. It is important that the details of pre-race data are accurate in order to maintain the distinctive qualities of betting on British horseracing (e.g. many punters get pleasure from studying information on a race before placing their bets). The current system provides an extremely high level of accuracy with very rapid internet dissemination of information which changes by the hour. It is difficult to envisage how this accuracy could be replicated by a third party (e.g. trainers might not inform all data providers of entries and withdrawals) and the loss of reputation for accuracy would be harmful all round. The problem is even deeper in that, in the absence of a body representing the collective interests of all those in racing, contributors to the database would likely claim ownership of their own bits of data (e.g. owners’ racing colours or entries, racecourses on race times, Jockey Club on handicaps) and the transaction costs of compilation would be enormous.

Conclusion on agreements that enable BHB to negotiate collectively BHB is a not-for-profit organisation that coordinates negotiations on behalf of non-bookmaker interests. Sixty per cent of the Levy goes towards prize money, thus cutting the price of racing for owners, and the remainder goes towards integrity services, developing racecourse facilities for the benefitof spectators and improving the breed. The Levy is a betting-profits tax and its proposed replacement was a charge for data based on betting profits. Neither is significantly distortionary for punters because it remains optimal for bookmakers to set the same odds that maximise their profits, whether they keep 100 per cent or only 90 per cent. For this and other reasons, British horseracing offers punters amongst the lowest price of a bet. The OFT proposals would have fragmented negotiations and undermined property rights, with the consequence of unbalancing the outcome. With very considerably reduced funding, the quality of British horseracing (the common C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 213 [192–216] 16.6.2009 11:53AM

213 8. The Orders and Rules of British Horseracing

product into which the various groups of consumers each buy) would be greatly diminished. Furthermore, competition in the betting industry would have been seriously undermined if individual negotiations had been required and resulted in exclusive deals.

Coordination of fixtures and race planning Another part of the OFT’s Statement of Objections (see section 4) related to ‘limiting the freedom of racecourses to organise their racing, in particular by fixing how often and at what times they stage races and the type of race they stage’. There are good reasons to have an element of central planning of fixtures. Different groups of consumers have somewhat different preferences over fixtures. For example, bookmakers want racing all day every day, spectators want weekend and Bank Holiday racing and owners need to plan a campaign for their horses over the season. The matching of the needs of the horse population and a suitable set of races, plus a relatively fixed pot of funding for prize money, cannot be achieved without an element of central planning for fixtures and race conditions. However, this raises the issue of which racecourses should be able to run which fixtures and whether competition could be beneficially introduced. Historically, fixture allocation was determined by ‘grandfather rights’, i.e. if you had a fixture on the first Friday in May last year, you have first rights on it this year. This had the inevitable consequence of freezing the number of courses and limiting the incentive for each course to work at developing sponsorship and facilities. Trade in fixtures between racecourses was very limited. Furthermore, there was a ‘fifty-mile rule’ which prevented two meet- ings from taking place on the same day at courses within fifty miles of each other. Some racecourses felt seriously constrained by these rules and lobbied hard for their reform, while others were disturbed at the prospect of change. While there are strong arguments for central planning of the general pattern of fixtures to meet the requirements of bookmakers and others, this could be achieved with far less rigid rules. In particular, the fifty-mile rule had little economic justification. Fixture allocation could also be freed up by an auction system for fixtures, with racecourses bidding for slots and revenues going towards overall prize money. In this respect, the OFT objections undoubtedly helped spur an emerging modernisation programme into action by leading to undertakings to abolish these unnecessary restrictions and to introduce rolling auctions for a subset of fixtures that had no great tradition (e.g. no auction of Derby Day which is traditionally run at Epsom). C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 214 [192–216] 16.6.2009 11:53AM

214 Cases in European Competition Policy: The Economic Analysis

Integrity and prize money The third OFT objection was that BHB should not restrict prize money for particular races. The OFT implied that prize money should be used in an unrestricted way by individual racecourses to attract horses to run, regardless of the quality of the race. This would allow higher prizes for races run by lower-grade horses than those rated more highly and so create severe incentive problems, undermining the integrity of races (e.g. an owner would be tempted to hold back a high-quality horse in some races in order to move down the ratings and so qualify for a low-grade but higher prize money race against weak opposition). Meritocratic prizes are particularly important for the integrity of handicap races. Handicapping is a major element in the mix of creating exciting races with attractive betting opportunities. It is essential to the integrity of such races that horses have the incentive to be run true and so provide information for handicapping on merit.44

6. Outcome and conclusion

This was a case that set some racecourses against the sport’s governing authority. It also set bookmakers against the rest in a classic fight over the allocation of the spoils derived from taking bets on British horseracing. Some very strong personalities and very old traditions were involved, using a very new weapon of competition law.45 Modernisation was overdue but a careful path needed treading if it was not to trample on what underpinned its success. In June 2004, four years after original notification, BHB reached agreement with the OFT.46 The OFT was ultimately persuaded to withdraw its preliminary view that it should fragment British horseracing. It withdrew its challenge to those Orders and Rules ‘monopolising’ the sale of pre-race data and allowing the governing body to coordinate key elements of the overall shape of fixtures and race planning across the season. It also accepted the need for

44 In fact, the modernisation programme proposed a more, not less, rigorous approach to linking prizes to race quality. Furthermore, unrestricted relative prize money would be impractical as a way of balancing race sizes (which was the supposed justification in the hypothesised ‘market for British racing opportunities’) because the supply of horses is very inelastic over the days or even hours necessary to fill a race, so the swings in prize money would have to be seriously large. 45 The UK Competition Act (1998) became operational from 1 March 2000 and introduced almost identical prohibitions on anticompetitive agreements (Chapter 1) and abuse of dominance (Chapter 2) to those in EC Articles 81 and 82 respectively. 46 OFT Case CP/1058/00 Notification of Governance Agreements 28 June 2004. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 215 [192–216] 16.6.2009 11:53AM

215 8. The Orders and Rules of British Horseracing

regulation of prize money in relation to race ratings. Two commitments were agreed in relation to the sale of pre-race data. First, licensing arrangements must be ‘on an open, non-discriminatory and non-exclusive basis, at a fair market price’. Second, in the event of pricing disputes, an independent ‘arbitrator will be required to resolve the dispute having regard to the needs of racing, the value of British horseracing to the bookmakers’ business and, importantly, competition law’. BHB also proposed to link the allocation of funding to the racecourses generating most betting revenue and to give them more discretion to set race programmes (i.e. incentivising courses to provide racing most attractive to punters). BHB further undertook to relax a number of other Orders that could not be justified, for example, competition between racecourses was enhanced by the abolition of the fifty-mile rule, the introduction of a set of new fixtures that BHB proposed to auction to racecourses on three-year ‘leases’.47 Minimum prize money for the lowest-grade races was also adjusted to facilitate more racing for less able racehorses. After a quarter of a millennium of institutional evolution, it is not surprising that reform was needed, and these agreed changes sat naturally as part of a much wider modernisation programme started earlier by BHB chairman Peter Savill. The wider economic lessons of this case include the need for a competition authority to address the efficiency motives for apparently anticompetitive agreements, the abolition of which may have severely negative side effects. There should be serious consideration of the counterfactual. It is also important to get the product definition right before rushing into multiple partial ‘market definitions’. Where different groups of consumers buy into the same product and this creates externalities, great care is needed to make sure these externalities are appropriately addressed. This has been a common theme of much recent economic research on what have become known as ‘two-sided markets’. Competition authorities should also beware of getting drawn inap- propriately into commercial disputes. With huge sums of money at stake, businesses will work through every available court using every available law to try to gain a larger slice of the cake, even when this reduces the size of the cake. The best way to limit the consequent inefficiencies is to focus clearly on economic effects and for economists to make the economic arguments under- stood by the competition agencies and courts.

47 This is an example of how bidding for the market can provide an alternative to more standard competition. This approach is familiar in the presence of natural monopoly (e.g. TV or rail franchises) where there are well-known dangers of ex-post compliance and renegotiation (Williamson, 1976). In the case of ﬁxture auctions, such dangers are minimised by the presence of alternative racecourses which could credibly take over a ﬁxture if the leaseholder tried to renegotiate. C:/ITOOLS/WMS/CUP/386783/WORKINGFOLDER/LYO/9780521886048C08.3D 216 [192–216] 16.6.2009 11:53AM

216 Cases in European Competition Policy: The Economic Analysis

Unfortunately, the legal issues in this case have been much harder to sort out. This case has links with a number of others which together touch on an enormous range of competition laws related to restrictive agreements, database rights, exploitative pricing, media rights and state aid. The courts have found the sale of pre-race data particularly confusing: is it excessive to negotiate a price of £100 million for something that costs only £5 million to compile? Or is £100 million a ‘fair price’ for something that provides access to a product costing many times that to create and which generates the buyer £1 billion in gross profits? Early court decisions and statements supported the BHB position (High Court, 2001; Advocate General of ECJ, July 2004). Later ones went against (ECJ, November 2004; UK Court of Appeal, July 2005), in particular undermining the rights over pre-race data. The last of these forced the UK Minister for Sport to agree that the HBLB must continue the Levy system until an alternative commercial mechanism can be found (though there appears to be no practical alternative to the sale of pre-race data). In a related case on the sale of pre-race data to Attheraces, a proposed television channel, the case was set up as an abuse of dominance by BHB, which lost on excessive pricing in the High Court (December 2005) before winning in the Court of Appeal (February 2007).48 All this has created huge uncertainty and delayed the modernisation programme. Non-specialist courts clearly find these issues very difficult, thus providing an object lesson in the value of the economic approach to competition analysis.

Acknowledgements

I am grateful to David Elliott and Steve Davies for detailed comments on this chapter. I acted as an expert witness on behalf of the BHB during the period of its response to the Rule 14 Notice (2003–4). I learned an enormous amount about the organisation of horseracing from Greg Nichols, then Chief Executive of BHB, with additional detail from other BHB executives and representatives of various horseracing bodies. Guy Leigh led the legal team. I worked closely with David Elliott with Eric Morrison at PwC. Much of the economic analysis in this chapter was jointly developed with David and Eric, to whom I am most grateful, though I am sure they will agree that any errors remain mine.

48 See Elliott (2007).